卫星ATM是近年来通信领域的研究热点 ,拥塞控制是其中很关键的一个问题 .本文主要研究基于星上处理的卫星ATM网的ABR流量控制 ,它通过对流量的长时预测来达到控制目的 .由于卫星信道的大延时 ,无法实现精确的长时预测 ,因而将预测结果直接用于ABR的拥塞控制效果不甚理想 .将长时预测同动态利用因子控制的方法结合之后 ,在链路效率、信元丢失率、不同连接数目和暂态响应等方面都取得很好的结果 . Satellite ATM is a research hotspot in the field of communication in recent years, and congestion control is one of the most important issues.This paper mainly studies the ABR traffic control based on satellite processing in satellite ATM network, which achieves the control purpose by long-term traffic prediction. Due to the large delay of satellite channel, accurate long-term prediction can not be achieved, so it is not effective to use the prediction result directly in the ABR congestion control.After combining the long-term prediction with the dynamic factor control, , Cell loss rate, the number of different connections and transient response have achieved very good results.